alainrk

Right, every technology is suboptimal in the beginning, but constraints will ever exists, energy and hardware in primis. Aldo you have to factor entropy in regardless, complexity remains a cost

Yeah but energy is still a cost, and local inference without batch and multiplexing for many users (like an office would be) is even less optimal. I'm not sure I think it's just pushing the problem forward

Indeed, but I also think structural considerations must be done at the company level on the work model overall

Interestingly hackernews commenters are the main reason of this fear

For context, this is an answer to a more pessimistic outlook here: https://mariozechner.at/posts/2026-03-25-thoughts-on-slowing...

Thank you for explaining that

Man it doesn't post automatically

Great analysis, you should open a blog with your great abilities

That's why LLM exist, you can go there, paste the link and specify the number of words you want the synthesis. Or just take a few minutes and read it entirely. Nobody forces you.

It is not name dropping, the last name is literally mentioned along with the name of the book.

How fun to see that the most common insult in 2026 is that something is AI-generated. Is this comment too? Can you provide the prompt you use and the automation you have to increase your karma points?

Sean Goedecke and many others do the same, is it ok in their case just because more famous?

Agree with this, I've been testing how AGENTS.md and similar can do to automatically have these behaviours and I feel (it's just feeling) it's been improving over time. Clearly depends a lot on the agent, the model, the codebase size and so on.

My complete reasoning, notes, errors have never been part of the commit. I don't see a valid reason on why the raw conversation must be included. Rather I have hooks (or just "manually" invoked) to process all of it and update the relevant documentation that I've been putting under docs/.

I can confirm, happened to me a few days ago

How did you find out?

Have you tried raptor? (https://github.com/gadievron/raptor)

Strawman argument.

Standards and patterns matter, but discernment matters more. The issue isn't reusability itself, it's the cargo-cult adoption of frameworks that solve problems you don't have, when you don't have them.

Your LLM agent works for undiscussed reasons because you made deliberate architectural choices for your specific context. That's engineering. Blindly importing a framework just because "everyone uses it" is the opposite. That's the point, nothing more nothing less.

I suggest to read the full article :)

If you have no idea on how to setup the pillars you're absolutely right, maybe you should try

Interesting analysis

Thank you the insightful feedback :) If you also have something to say on the point of the article itself, instead of pointing the finger on the person I'll be happy to answer on that

I agree with your point. My concern is more about the tedious aspects. You could argue that tedium is part of what makes the craft valuable, and there's truth to that. But it comes down to trade-offs, what could I accomplish with that saved time, and would I get more value from those other pursuits?

Indeed, this has been one of the first things I've noticed

Thank you for the constructive feedback :)

You're right, clearly I've tried to be a bit provocative to pass the message, but I'm not religious in this sense. Minimal frameworks that really solve a problem cleanly and are adopted with intention are welcome.

Thank you, I'm glad to be back!

Totally agree with this!